Nvidia is preparing to launch the fifth iteration of its Deep Learning Super Sampling (DLSS) technology next month, starting with an integration into Hogwarts Legacy. Following an initial reveal earlier this year that drew significant criticism for its aggressive use of generative artificial intelligence, the company has reportedly decided to scale back the technology’s impact on character visuals. According to reporting from The Verge, the version of DLSS 5 appearing in the game will be toned down compared to previous demonstrations that many users found visually disruptive.
The Shift Toward Generative Upscaling
DLSS has evolved from a spatial upscaler into a complex suite of AI-driven tools designed to improve performance and image quality. While previous versions focused on increasing frame rates through frame generation and improving ray-traced lighting, DLSS 5 represents a shift toward using generative AI to alter or enhance game assets in real-time. This approach initially faced backlash from the gaming community. Early previews showed the technology significantly altering character models, such as making a teenage student in Hogwarts Legacy appear noticeably older or changing the facial features of characters in titles like Resident Evil. Critics and players argued that these AI-driven transformations compromised the artistic intent of the game developers by introducing unexpected visual hallucinations.
Technical Implementation and Requirements
The announcement did not specify the exact technical methods Nvidia is using to tone down the generative aspects of DLSS 5 for its debut. It remains unclear if the software will offer a user-controlled slider for AI intensity or if certain generative features have been disabled entirely for this specific release. Furthermore, the reporting did not provide a specific launch date within next month, nor did it clarify which Nvidia RTX graphics cards will be required to run the new software. The impact on system performance—specifically whether this generative processing requires more hardware overhead than the current DLSS 3.5 standard—is also currently unknown.
As DLSS 5 prepares for its first major integration, the industry will be watching to see if Nvidia can balance the performance benefits of AI with the need for visual accuracy. The success of this rollout may determine how generative AI is utilized in future game engines.
Lead image: Matheus Bertelli / Pexels
